Measure the arm of a Carnotaurus, and you get a number that feels wrong. The Carnotaurus was a bull-horned predator from Patagonia, roughly the length of a bus, built for violence. Its arms were shorter than its skull. Shorter, in fact, than those of Tyrannosaurus rex, which itself had arms so reduced that palaeontologists have spent over a century debating what they were actually for. Now a study from UCL and Cambridge may have an answer, and it turns the question on its head: the arms didn’t wither in isolation. The skull, it seems, ate them.
The research, published in Proceedings of the Royal Society B, examined 82 species of theropod dinosaur, the two-legged mostly carnivorous group that includes T. rex, Velociraptor and, distantly, every bird alive today. What the team found was a statistical signal strong enough to outlast every control they applied: across at least five separate lineages, forelimb reduction was tightly coupled with cranial robusticity, a composite measure of how powerful and biomechanically formidable a skull had become.
Lead author Charlie Roger Scherer, a PhD student at UCL Earth Sciences, put it plainly. “Everyone knows the T. rex had tiny arms but other giant theropod dinosaurs also evolved relatively small forelimbs,” he said. “The Carnotaurus had ridiculously tiny arms, smaller than the T. rex.” The pattern had been noticed before in individual lineages but had never been rigorously tested across all of Theropoda simultaneously. Scherer’s team built something they called the Cranial Robusticity Score, incorporating skull proportions, bone fusion, dental morphology and estimated bite force, then ran it against forelimb data from dozens of species. The correlation came out clearly: the more reinforced the skull, the more reduced the arms.
An Arms Race Built Around Jaws
The team’s explanation runs roughly as follows. Across the Mesozoic, prey available to large carnivorous dinosaurs kept getting bigger. Sauropods, the long-necked plant-eaters that shared these ecosystems, reached sizes still difficult to fully comprehend. Hunting animals that large required a different toolkit. Claws and grasping forelimbs, useful enough when wrestling prey of comparable size, become somewhat beside the point when your quarry is a creature ten times your mass.
Jaws, on the other hand, scale rather well. A sufficiently robust skull, with bones fused tightly and jaw muscles enlarged to a grotesque degree, can deliver bite forces capable of restraining and dismembering enormous prey. Scherer described it as a case of “use it or lose it.” “The head took over from the arms as the method of attack,” he said, and as hunting shifted toward jaw-dominated strategies, the forelimbs gradually became redundant. Structures that aren’t being used tend to shrink; maintaining them carries a metabolic cost that natural selection will eventually prune away.
That sequence matters. The team’s ancestral state reconstructions, which track trait changes across an evolutionary tree calibrated against the fossil record, consistently showed skull robusticity and body mass increasing before forelimb reduction appeared, never after. “It would not make evolutionary sense for it to occur the other way round,” Scherer said, “and for these predators to give up their attack mechanism without having a back-up.” A statistical test used to assess causal direction between variables pointed the same way: forelimb reduction appears to be a downstream consequence of cranial evolution.
Five Roads to the Same Diminished Destination
One of the more striking findings is that tiny arms evolved independently in at least five theropod lineages: abelisaurids, carcharodontosaurids, ceratosaurids, megalosaurids and tyrannosaurids. They were not closely related, and they weren’t solving the problem identically. In abelisaurids such as Carnotaurus, reduction appears to have started in the hands and lower arm. In tyrannosaurids, every part of the forelimb shrank at roughly similar rates. Same outcome, different developmental path.
The Majungasaurus is worth pausing on. An apex predator in what is now Madagascar around 70 million years ago, it weighed only about 1.6 tonnes, a fifth of T. rex’s bulk. Its skull was extraordinarily robust for its size, and its forelimbs were proportionally among the most reduced of any known predatory dinosaur. Which rather undercuts the simpler version of the story, the one where tiny arms are just what happens when a dinosaur gets very large. Body size mattered, the study found, but chiefly as a secondary factor. Cranial robusticity was doing the heavy statistical lifting.
Some predators didn’t follow the pattern. Spinosaurids grew to comparable sizes but retained elongated forelimbs, apparently because they hunted fish rather than megaherbivores and had little incentive to swap grasping limbs for a bite-dominated strategy. “Trying to pull and grab at a 100ft-long sauropod with your claws is not ideal,” Scherer observed. “Attacking and holding on with the jaws might have been more effective.” Ecological context shaped the trajectory; gigantic prey was probably necessary, but not quite sufficient on its own.
There is, as always, a caveat. The team is transparent about it: the study identifies a pattern and proposes a mechanism consistent with it, but correlation isn’t causation, and the fossil record doesn’t come with timestamps fine-grained enough to rule out every alternative. What the research does resolve, at least in broad terms, is one of palaeontology’s more beloved puzzles. Various functions for the arms have been proposed over the decades, from use in mating to helping a fallen animal rise. Those hypotheses now look less like explanations and more like post-hoc attempts to find a role for something that evolution had already effectively abandoned.
Five separate lineages of dinosaurs arrived at much the same answer, each by a slightly different route. The answer was: less arm, more head. That the world’s most iconic predator ended up as supporting evidence for a principle of evolutionary economy, its famous tiny limbs recast as metabolic leftovers rather than mysterious specializations, is the kind of resolution that palaeontology occasionally delivers. Tidier than the mystery. Somewhat harder to put on a museum placard.
Frequently Asked Questions
Why did T. rex have such small arms?
According to a new study, the arms shrank because the skull took over as the primary weapon. As T. rex and its relatives evolved increasingly powerful, heavily reinforced jaws capable of subduing enormous prey, the forelimbs became less important to hunting and gradually reduced over millions of years. Skull robusticity turns out to be the strongest statistical predictor of arm reduction across theropod dinosaurs, stronger even than overall body size.
Was T. rex the dinosaur with the smallest arms?
Not quite. The Carnotaurus, a horned predator from Patagonia, had proportionally smaller arms than T. rex relative to its skull. The Majungasaurus, a smaller predator from Madagascar weighing around 1.6 tonnes, also had extremely reduced forelimbs, showing that tiny arms were not simply a by-product of growing very large.
Did other dinosaurs also evolve small arms?
Yes, and this is one of the study’s key findings. At least five separate dinosaur lineages independently evolved reduced forelimbs: abelisaurids, carcharodontosaurids, ceratosaurids, megalosaurids and tyrannosaurids. They were not closely related, which means the same outcome evolved multiple times in response to similar ecological pressures, particularly the challenge of hunting very large prey.
Why didn’t all large predatory dinosaurs evolve tiny arms?
Ecological diet played a significant role. Spinosaurid dinosaurs grew to comparable sizes but retained elongated forelimbs, apparently because they hunted fish and other smaller prey rather than enormous sauropods. The shift toward jaw-dominated hunting, driven by increasingly gigantic herbivore prey, seems to have been the critical factor; predators that didn’t face the same prey demands kept their arms.
Source: Scherer CR, Steell E, Upchurch P. “Drivers and mechanisms of convergent forelimb reduction in non-avian theropod dinosaurs.” Proceedings of the Royal Society B (2026). https://doi.org/10.1098/rspb.2026.0528
ScienceBlog.com has no paywalls, no sponsored content, and no agenda beyond getting the science right. Every story here is written to inform, not to impress an advertiser or push a point of view.
Good science journalism takes time — reading the papers, checking the claims, finding researchers who can put findings in context. We do that work because we think it matters.
If you find this site useful, consider supporting it with a donation. Even a few dollars a month helps keep the coverage independent and free for everyone.